Fast-Fourier transform

[rainny]

Hi, does anybody know how to apply the fast-Fourier transform on
spectrum analysis?
When the application play the sound, it got the frequency wave shown,
while if there are no sound, the frequency wave will not be shown.

Does anybody have any idea on it? I try to search info about the fast-
fourier transform but unable to find the correct information.

If anybody knows about it, please help me.

Regards
Rainny

 

[softwarepearls_com]

Hi, does anybody know how to apply the fast-Fourier transform on
spectrum analysis?
When the application play the sound, it got the frequency wave shown,
while if there are no sound, the frequency wave will not be shown.

Does anybody have any idea on it? I try to search info about the fast-
fourier transform but unable to find the correct information.

If anybody knows about it, please help me.

Regards
Rainny

The way I would tackle this is to read the Wikipedia entry/entries on
this subject, then search for C code which does what you want, and
then maybe port the logic to Java.. Also check if there's no Java
library already out there..

 

[rainny]

Yes, I also search the wiki about the fast fourier transform, but the
information only shows about the algorithm. But, what I want is apply
the algorithm into speech and signal processing. So far, I can't find
the solution on it..
Is there anybody still have any idea on it?

Thanks and regards
rainny

 

[Gordon Beaton]

Yes, I also search the wiki about the fast fourier transform, but the
information only shows about the algorithm. But, what I want is apply
the algorithm into speech and signal processing. So far, I can't find
the solution on it..
Is there anybody still have any idea on it?

Several of these code libraries can do FFT:
http://math.nist.gov/javanumerics/

/gordon

--

 

[Bo Vance]

Hi, does anybody know how to apply the fast-Fourier transform on
spectrum analysis?
When the application play the sound, it got the frequency wave shown,
while if there are no sound, the frequency wave will not be shown.

Does anybody have any idea on it? I try to search info about the fast-
fourier transform but unable to find the correct information.

If anybody knows about it, please help me.

Regards
Rainny

In addition to the links I posted several days ago on this same topic.

The Scientist and Engineer's Guide to Digital Signal Processing
<http://www.dspguide.com/>

 

[rainny]

Actually, I choose this algorithm because the algorithm mention that
can apply in the speech and signal processing. Thus, what I really
search for is use the algorithm to process the incoming audio. When
indication of the sound happens, then the application will show the
frequency wave.

 

[Roedy Green]

When the application play the sound, it got the frequency wave shown,
while if there are no sound, the frequency wave will not be shown.

A sound is a change in air pressure at a regular rate. the Frequency
is how fast it changes. If there is no sound, there in so change in
air pressure, so it makes to sense to ask at what frequency it is
changing.

Googling on FFT and Fast Fourier Transform will probably find you lots
of reading material.

My own introduction to FFT was quite fun. I did FFTs on dolphin
sounds with special purpose hardware. Software was not fast enough at
the time to keep up with real time.

 

[Patricia Shanahan]

you know that the purpose of FFT is to multiply polynomes in O(n log n)
instead of O (n^2) ?
I am sure there are many applications to polynome multiplication in
signal processing. Though ... its not like you just throw FFT at the
audiostream and get what ever you want out of it... FFT is no golden
hammer.

Huh? I thought that the purpose of FFT is to translate wave form data,
in either direction, between the time domain and the frequency domain.
Time domain data is represented by samples at various times. Frequency
domain data is represented by a frequency spectrum.

FFT is important because it has many, many practical applications in
signal processing. Some calculations are much easier in one domain than
the other.

Its application to multiplication is an interesting demonstration of
its versatility, but not particularly relevant to audio signal
processing. The basic objective of translation between time domain and
frequency domain is much more relevant.

Patricia

 

[John B. Matthews]

Actually, I choose this algorithm because the algorithm mention
that can apply in the speech and signal processing. Thus, what I
really search for is use the algorithm to process the incoming
audio. When indication of the sound happens, then the application
will show the frequency wave.

[Y]ou know that the purpose of FFT is to multiply polynomes in O(n
log n) instead of O (n^2) ? I am sure there are many applications
to polynome multiplication in signal processing. Though ... its not
like you just throw FFT at the audiostream and get what ever you
want out of it. FFT is no golden hammer.

Huh? I thought that the purpose of FFT is to translate wave form data,
in either direction, between the time domain and the frequency domain.
Time domain data is represented by samples at various times. Frequency
domain data is represented by a frequency spectrum.

That's my understanding. I thought Christian was simply referring to the
convolution theorem:

FFT is important because it has many, many practical applications in
signal processing. Some calculations are much easier in one domain than
the other.

Indeed, FFT (and Moore's law) has dramatically improved the speed of
medical imaging, for example:

<http://en.wikibooks.org/wiki/Basic_Physics_of_Nuclear_Medicine/Fourier_M
ethods>
<http://en.wikibooks.org/wiki/Basic_Physics_of_Nuclear_Medicine/Deconvolu
tion_Analysis>

Its application to multiplication is an interesting demonstration of
its versatility, but not particularly relevant to audio signal
processing. The basic objective of translation between time domain
and frequency domain is much more relevant.

Surely.