A fast algorithm of running fourier transform
Hiroyuki Kamei, Tetsuya Harada, Hiroshi Kawarada · Electronics and Communications in Japan (Part I Communications) · 1988
Abstract A short‐time Fourier transform can be derived by low‐pass filtering of the product of an input signal and exp(j2°ft). According to this method, called the Running Fourier Transform (RFT) in this paper, running power spectra with arbitrary center frequencies and arbitrary Q values can be obtained. This paper proposes a fast algorithm of discrete RFT (FRFT). In the FRFT, a first‐order lag system is adopted as the low‐pass filter (LPF), and by approximating the impulse response of the LPF with a step function, the amount of multiplications is reduced. The calculation of the complex exponentials is omitted by referring to a table of sin(2°k/K) (k =0, …, K).