南京理工大学数字信号处理英文课件 Lecture19.ppt

261 * Lecture 19 Discrete Fourier Transform( II ) Discrete Fourier Transform (DFT) ? Definition ? Properties ? A finite-length sequence x[n] can be represented by samples of its DTFT spaced by 2π/N, where N ≥length(x[n]) these samples are the DFS sequence of a periodized version of x[n] If we limit the length of X[k] to one period, we obtain the DFT of x[n]: . DFT: Introduction Means that: ---- Analysis equation ---- Synthesis equation A. DFT: Definition otherwise otherwise Let a sequence x[n] have length N: Example: a finite-length sequence: DTFT: DFT: Example x[n] X1[k] |X2[k]| Arg{X2[k]} DFT: Example ( II ) B.1 DFT: Properties ? linearity: ax[n] + by[n] DFT aX[k] + bY [k], where x[n] and y[n] have the same length (after zero-adding if necessary) ? Shift: the sequence that has W kmN X[k] as DFT ? Let Y [k] = W kmN X[k], then so that and i.e. a circular shift Example: circular shift B.2 DFT Properties: Circular Shift B.2 DFT Properties: circular shift ( II ) by using the duality property of the DFS, one proves that i.e. – periodize x[n] – flip the horizontal axis – keep one period from 0 to N－ 1 B.3 DFT Properties: Duality B.3 DFT Properties: Duality ( II ) x[n],N=10 Example: B.3 DFT Properties: Duality ( III ) Then: and B.4 DFT Properties: Symmetry ? periodic conjugate-symmetric and conjugate-antisymmetric parts: * * *