FFT Analyzer Basic FAQ

The input time signal is generally a continuous signal that continues indefinitely. FFT processing analyzes this signal by cutting it out at a time window length T (number of sampling points N). Since FFT calculation processes the real signal by cutting it out at a finite time length T and treating it as a repeating time waveform, when T is an integer multiple of the input signal period, the repeating waveform is continuous and a correct spectrum is obtained. However, if this is not the case, the waveform becomes discontinuous before and after the repeating point, causing distortion in the waveform, and the spectrum spreads out around that frequency. As a result, the magnitude of the peaks in the spectrum decreases compared to the true value, and that amount of power (amount equivalent to magnitude) leaks out on both sides of the peak.

This process of extracting a finite time segment T from a real signal is called applying a time window, and the resulting spread of the spectrum is called leakage error. Such errors can be a significant practical problem because high-power frequency components can mask nearby small peaks, so efforts are made to minimize them as much as possible.