2 2dft
2 2dft
2 2dft
N
X −1
X(k) , DFT[x(n)] = x(n)WNkn (1)
n=0
0≤k ≤N −1
− 2π
donde WN , exp N .
N −1
1 X
x(n) , IDFT[X(k)] = X(k)WN−kn (2)
N
k=0
0≤n≤N −1
function Xk=dft(xn)
% X(k)=DFT(x(n))
%
% Calcula la DFT de la secuencia x(n) de duracion finita N.
N=length(xn);
n=0:N-1;
k=0:N-1;
WN=exp(-j*(2*pi/N));
W=WN.^(k’*n);
Xk=W*xn’;
function x=idft(X)
% x(nk)=IDFT(X(k))
%
% Calcula la IDFT de la secuencia X(k) de duracion finita N.
N=length(X);
n=0:N-1;
k=0:N-1;
WN=exp(j*(2*pi/N));
W=WN.^(k’*n);
x=(1/N)*W*X’;
N
X −1
X(k)
e = e(n)WNkn,
x 0≤k ≤N −1 (3)
n=0
N −1
1 X e
x
e(n) = X(k)WN−kn, −∞ ≤ n ≤ ∞ (4)
N
k=0
2. La DFT en cambio:
N
X −1
X(k) = x(n)WNkn, 0≤k ≤N −1 (5)
n=0
2) Ecuación de sı́ntesis:
N −1
1 X
x(n) = X(k)WN−kn, 0≤n≤N −1 (6)
N
k=0
3. Comparación resumida:
DFS DFT
−1
NP −1
NP
X(k)
e = e(n)WNkn
x X(k) = x(n)WNkn
n=0 n=0
0≤k ≤N −1 0≤k ≤N −1
−1
NP −1
NP
−kn
x
e(n) = 1
N X(k)W
e
N x(n) = 1
N X(k)WN−kn
k=0 k=0
−∞ ≤ n ≤ ∞ 0≤n≤N −1
o equivalentemente:
x
e(n) = x(nmodN ), −∞ ≤ n ≤ ∞ (8)
c) Ecuación de sı́ntesis:
Z π
1
x(n) = X(ω)eωn dω (11)
2π −π
n0 +(N −1)
X
X(ω) = x(n)e−ωn (12)
n=n0
y si x(n) es causal:
N
X −1
X(ω) = x(n)e−ωn (13)
n=0
0≤k ≤N −1
donde
sin ωN
2 −ω ( N 2−1 )
Φ(ω) = ω e
N sin 2
−1
NP −1
NP
− 2π
N kn
X(ωk ) = x(n)e X(k) = x(n)WNkn
n=0 n=0
0≤k ≤N −1 0≤k ≤N −1
− 2π
9. Tomando en cuenta que WN = exp N concluimos que:
0≤k ≤N −1
Zero-padding
X(k) = X(ω)|ω= 2π k 0≤k ≤N −1
N
Dada la señal
xa(t) = cos ω1t + cos ω2t
6. Comentar y concluir.
Propiedades de la DFT
1. Linealidad:
2. Rotación circular:
define como:
x(0), n = 0;
x((−n))N =
x(N − n), 1 ≤ n ≤ N − 1.
DFT[x∗(n)] = X ∗((−k))N
X(k) = X ∗((−k))N
DFT [x1(n)
N x2 (n)] = X1 (k)X2 (k)
100
90
80
70
60 DFT(N2)
50
40
30
20
FFT(N logN)
10
0
1 2 3 4 5 6 7 8 9 10
N
Referencias
[1] Vinay K. Ingle and John G. Proakis. Digital Signal Processing us-
ing MATLAB. BookWare Companion Series. Brooks/Cole Thomson
Learning, 2000.
[4] Gordon E. Carlson. Signal and Linear System Analysis with MATLAB.
BookWare Companion Series. John Wiley & Sons, Inc., 1998.