Narrow Band Spectrum

A narrow band spectrum is a spectrum calculated using a Fast Fourier Transform (FFT), and hence provides an analysis of the signal energy within narrow frequency bands (the width of these bands depends on the chosen length of the FFT).

Several types of narrow band spectrum can be calculated and displayed in SAS. These are introduced below.

Usually, a spectrum calculation starts with cutting the signal into several slices of a same length (Window size), multiplying each slice by an analysis window (of Window type), and then computing a FFT (of length FFT size) for each slice of the signal. The slices of the signal can overlap.

The following parameters are used in spectrum calculations:
  • Window type: the type of analysis window to use for the FFT analysis. SAS proposes a set of standard windows (Blackman, Flattop, Gauss, Hamming, Hann, Rectangular, Triangular).
  • Window size: this is the number of samples of the signal to be used for each FFT calculation. The analysis window size cannot be greater than the FFT size.
  • FFT size: this is the number of samples on which the Fast Fourier Transform (FFT) is performed. The higher the number of samples, the higher the precision in frequency. This parameter corresponds to the number of narrow bands used for the FFT analysis.
  • Overlap: this value specifies the overlapping ratio (in percent) between two successive signal slices on which the spectrum is calculated. The higher the ratio, the higher the number of signal slices.
Note: When the Window size is smaller than the FFT size, zero-padding is applied. This means that the analysis window is applied to Window size samples of the signal, and then a number of zero samples are appended, to match the size of the FFT samples. The FFT is then calculated on this "zero-padded" vector of samples.
Tip: Select the Same as FFT size option to use the same number of samples for the FFT and the Analysis Window. This is the recommended setting.

From this same method of calculation, several types of spectrum can be derived. These are described in the sections that follow.