Basic FAQ for FFT Analyzer

The average of the spectrum can be divided into two types: [1] power spectrum average and [2] Fourier spectrum average.
The main purpose of performing the average of the power spectrum in 1ch measurement is to average the fluctuating spectral values if it is a deterministic signal (for example, a periodic signal), and to improve the statistical spectral estimation accuracy for irregular signals. The averaging methods include: [1] summation average, [2] exponential average, [3] peak average, and [4] sweep average.

Summation average is a simple average (ensemble average) of power values for each frequency band, also known as RMS average. Exponential average is a weighted power average where the number of averages corresponds to the time constant. Peak average is a method that retains the maximum power value in the frequency band (for an average count of N times) and is not strictly an averaging process. Sweep average is a method that uses a sine wave signal and sweeps (i.e., sequentially changes the frequency) to obtain the spectrum, which is divided into internal signal mode and external signal mode. In the internal signal mode, a built-in oscillator equipped in the FFT analyzer is used, and a sine wave fully synchronized with the resolution of the FFT is used, allowing for accurate measurements.

What to be cautious about in power spectrum averaging is that it is an averaging process based solely on power values (i.e., EU ² values). For specific averaging methods, please refer to the related items.

Fourier spectrum average is an average over complex numbers, and it has the same effect as time waveform averaging when used in conjunction with a trigger function for synchronous averaging. Although it has the disadvantage of requiring a trigger signal, it has advantages over power spectrum averaging, including [1] reduction of noise components and [2] averaging of phase spectrum.

	Averaging Process (Basic Terminology for FFT)