基于Python实现快速傅里叶变换FFT

hjc2020

已于 2023-12-02 16:44:53 修改

阅读量478

点赞数 10

文章标签： python 算法信号处理

于 2023-12-02 16:42:18 首次发布

本文链接：https://blog.csdn.net/qq_43398777/article/details/134754346

版权

def fft(x, flag=True):

    """
    flag is True by default to represent the forward FFT
    flag is False, if inverse FFT is wanted
    this is a simple version for understanding FFT quickly
    the length of x should be even to be correct
    """
    
    n = len(x)
    factor = 1 if flag else 0.5
    if n == 1:
        return x
    w = np.exp(2 * np.pi * 1j/ n) if flag else  np.exp(-2 * np.pi * 1j / n)
    x_even, x_odd = x[::2], x[1::2]
    y_even, y_odd = fft(x_even, flag), fft(x_odd, flag)
    y = np.array([0 + 0j] * n)
    for i in range(int(n/2)):
        y[i] = y_even[i] + w**i * y_odd[i]
        y[i + int(n/2)] = y_even[i] - w**i * y_odd[i]
    return y * factor