Scribd
Upload
7 views1 page

Mixed Radix FFT

The document presents the derivation of the Mixed-Radix Decimation-in-Time Fast Fourier Transform (FFT) starting from the definition of the Discrete Fourier Transform (DFT) for a composite length sequence. It details the reordering of summation, splitting of complex exponentials, and the concept of time-decimated sequences. The final result shows how the DFT of size mp can be constructed from p DFTs of size m.

Uploaded by

woxawi6937
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
7 views1 page

Mixed Radix FFT

The document presents the derivation of the Mixed-Radix Decimation-in-Time Fast Fourier Transform (FFT) starting from the definition of the Discrete Fourier Transform (DFT) for a composite length sequence. It details the reordering of summation, splitting of complex exponentials, and the concept of time-decimated sequences. The final result shows how the DFT of size mp can be constructed from p DFTs of size m.

Uploaded by

woxawi6937
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Derivation of the Mixed-Radix Decimation-in-Time Fast Fourier

Transform
Document placed in the public domain by Mark Borgerding,2003

Start with the definition of the Discrete Fourier Transform F for the time sequence f. The
length of f is a composite number N=mp.
N −1 2 n k
−j
F N k , f = ∑ f ne N

n=0

Reorder the summation as a double summation.


m−1 p−1 2 k qpr 
−j
F m p k , f = ∑ ∑ f qpr e N

q=0 r=0

Split the complex exponential into two parts.


m−1 p−1 2 k qp 2 k r
−j −j
F m p k , f = ∑ ∑ f qpr e N
e N

q=0 r=0

Pull one of the exponents out of the inner summation.


p−1 2 k r m−1 2 k qp
−j −j
F m p k , f = ∑ e N
∑ f qpr e N

r=0 q=0

Factor out p from the inner exponent. Notice how similar the inner summation looks to
the original definition above.
p−1 2 k r m−1 2 qk
−j −j
F m p k , f = ∑ e N
∑ f qpr e m

r=0 q=0

Observe the time sequence f(qp+r) is a decimation of the original sequence. This is
where the term Decimation in Time (DIT) originates.

Let f r be the time-decimated sequence, such that f r q= f qpr  .

The symmetry of complex exponentials tells us that


F N k , f =F N  k mod N , f 

In this way, the DFT of size mp can be constructed from p DFTs of size m.
p−1 2 k r
−j
F m p k , f = ∑ e N
F m k mod m , f r 
r=0

You might also like