数字信号处理双语版ppt第五章.ppt

The basic butterfly unit is shown in the following Figure Exploiting the properties of the complex exponential sequence : (1) Complex Conjugate Symmetry: (2) Periodicity in Arguments: In-Place Computations Decimation-in-Frequency FFT Algorithmwith Radix-2 Decimation-in-Frequency FFT Algorithmwith Radix-2 Decimation-in-Frequency FFT Algorithmwith Radix-2 Decimation-in-Frequency FFT Algorithmwith Radix-2  Decimation-in-Frequency FFT Algorithmwith Radix-2  Decimation-in-Frequency FFT Algorithmwith Radix-2  Linear Convolutions Evaluated by the Circular Convolution Linear Convolutions Evaluated by the Circular Convolution Linear Convolutions Evaluated by the Circular Convolution Linear Convolutions Evaluated by the Circular Convolution Linear Convolutions Evaluated by the Circular Convolution Linear Convolutions Evaluated by the Circular Convolution Linear Convolutions Evaluated by the Circular Convolution Direct Method The Overlap-Add Method  Computational Method of the Inverse FFT * Table 5-1 Input Index Bit Reversal for an 8-point FFT ▼ What is the in-place computations? (1) Stage-by-stage, starting at stage and ending at stage . (2) Butterfly-by-butterfly in every stage; (3)An array used to store complex data is sufficient to store output data for all stages. Such a computational structure is the so-called in-place computations of the decimation-in-time FFT algorithm. 1. Decomposing the DFT into Two DFTs. Considering and noting , we have Substituting into the above equation, respectively, we get an even- numbered subsequence and an odd-numbered subsequence Let and Then, and Two sequences can be obtained by the following butterfly unit: After the first decomposition, the computational structure